tags: music mus-407 electroacoustic computer digital audio sampling sample-rate bit-depth sample dynamic-range quantization

Sampling (Signal Processing)

Sampling involves taking periodic measurements of an analog [signal] and assigning a digital value to each measurement, facilitating the conversion from analog audio to [digital audio].

• #sampling-signal-processing
  • #concepts
  • #sources

Concepts

• [Sample Rate]
• [Bit Depth]
• [Nyquist Frequency]
• [aliasing]
• [quantization]
• [Quantization error] and [Bit depth]
  • In general: greater bit depth means higher sample resolution, which means smaller quantization error, leading to better signal-to-[noise] ratio
• [dither]

To convert an analog signal to a digital signal, we take a periodic sample of the analog signal. The value we obtain is the [amplitude] of the signal at a specific point in time, corresponding to the pressure change captured by some device.

We take a new sample at a specific rate, called the [sample rate]

The amplitude measurement must be represented in a digital format through a process called [quantization]. Since digital values can only be expressed finitely, we must assign a range of values to work with to represent amplitude; we call this range [bit depth].

The difference between the original amplitude measurement and our new, quantized value according to the bit depth is called [quantization error], which is heard as [noise]. We can mitigate this noise through [dither]ing.

Due to the finite number of samples we can take in a given point in time, there is a limited number of high [frequencies] we can digitally represent. This limit is determined by the [Nyquist frequency].

Any frequencies above that limit causes [aliasing].

Sources

• MUS 407 Sampling